by Kory Stamper
You’ll notice a letter missing from that litany, and that is because it merits special mention. S is, to put it in the modern vernacular, the worst. It is the longest letter in the book and an absolute heartbreaker, because you can see the end of the alphabet from it, and you know that once you clear S, you are moving on to T–Z, and half of those are barely even letters. But S—S goes on for-fucking-ever. Exactly 11 percent of your dictionary is made of words that begin with S. One-tenth of your dictionary is made up of one twenty-sixth of the alphabet. I bet the guy in the picture who supposedly went home and shot himself was in the middle of S when he did.
It’s not just the length of the letter but the content as well that can get to you. Emily doesn’t mind S as much as she does D, because D is long and filled with horrible words (like “despair,” “dismal,” “death,” and “dejected”). “It’s depressing,” she said. “And even that starts with d!” I am not fond of G because it contains “get,” “give,” and “go,” and those are horrible entries,*6 but I love J because it is short and has “jackass” and all its sibling entries in it (including not only “jackassery” and “jackassness” but “jackass bark,” “jackass bat,” “jackass brig,” “jackass clover,” “jackass deer,” “jackass fish,” “jackass hare,” “jackass kingfisher,” “jackass penguin,” “jackass rabbit,” and “jackass rig”).
There are occasions, however, when even S holds a secret delight or two in the form of a pink, such as the one I found among the citations (or “cits” for short) for “sex kitten”:
sex kitten
sex pot
There is no essential difference in these defs [definitions], but they’re not the same. Some differentiation shd be made.
The pink was written by one of our former physical science editors infamous for commenting as brusquely as possible on things beyond his remit. The files are littered with his pinks, most of them gruffly correcting a defect that he’s perceived (correctly or otherwise) with an existing definition—any existing definition, not just one touching on the sciences. This is one of his more frustrating pinks. If he thought that the definitions could be differentiated, he clearly had some idea how while writing the note but decided to withhold that information.
Another one of our science editors who was reviewing the batch later was apparently irritated by this note, and decided to comment on what he no doubt saw as needless meddling. His typewritten response to the note about “sex kitten” reads, “I will no doubt regret saying this but I think you have misconstrued the meaning of ‘physical’ science somewhere along here.”
But a pink’s a pink. Steve acted on it for the Tenth, adding the word “young” to the definition for “sex kitten.”
—
Lexical definitions come in a handful of philosophical-sounding flavors—ostensive, synonymous, analytical, truncated, periphrastic. These are fancy names that describe, in order of attempt, what each native English speaker does or attempts to do when called upon to explain what a word means. Everyone everywhere can do ostensive defining: it’s the act of physically showing a person what the word means. When a beaming, sleep-deprived parent asks his infant, “Where’s your nose?” and the baby blops a fist onto the middle of her face, that is ostensive defining.
Most dictionaries avoid ostensive defining because it’s hard to do ostensive defining with anything remotely abstract: How do you point at “sad” or “concept” or “for”? Ostensive defining also limits you to one example of a type: cups come in a wide variety of shapes, sizes, colors, and materials, but you may only have a red plastic disposable cup handy when someone asks, “What’s a cup?” We understand that there’s wide variety in some things (cups) and not in others (okapis), but dictionaries are not set up to adapt to a user’s mental pictorial database of things. That said, we do indulge in adding ostensive defining to our dictionaries from time to time in the form of pictures or illustrations. We trust our users to be able to read the definition of “gable”—“the vertical triangular end of a building from cornice or eaves to ridge”—and figure out what part of the building that definition refers to, but sometimes it’s easier to provide an illustration of a roof with a line pointing to the gable. “Picture dictionaries” or “visual dictionaries,” which are usually geared toward children who can’t read yet or for English-language learners, make heavy use of ostensive defining.
As with ostensive defining, nearly everyone can come up with synonymous definitions: they are simply definitions that list other better-known synonyms of a word. Your great-auntie Rose calls someone a “schlemiel,” and when you ask what “schlemiel” means, she tells you it means “idiot” or “fool” or “dupe” or “chump” ad infinitum.*7 You ask what “beautiful” means, and someone responds with “pretty.” This sort of defining gets ingrained in us through a collective childhood of vocabulary quizzes and standardized tests. Dictionaries do quite a bit of synonymous defining because we claim it makes semantic relationships between different types of words apparent. It also just so happens to be a lot easier to read a one-word synonym than to decipher a lengthy definition.
The jump from synonymous defining to analytical defining is a big one. Analytical definitions are the most common ones you’ll find in dictionaries, the ones that read like they were written by a team of neurodiverse robots—the ones that take years of practice to write.
A definition begins with what we call the “genus,” the overarching category that describes what, at heart, the “definiendum,” or word being defined, means.*8 The genus term must be broad enough to encompass the wide variety of uses for that word, but not so broad that it provides inadequate direction. Definienda should be, our training documents say, placed in the smallest genus that can contain them, lest they rattle around in there and irritate you. Sometimes this is easy: “snickerdoodle” is clearly “a cookie” and not “a dessert” (which lumps it in with pudding, pie, and ice cream) or “a meal” (which makes it an appetizer or entrée). More often, it’s not as simple. What genus term works best for “surfboard”? “A piece of sporting equipment,” perhaps? No: the genus in that phrase would be “piece,” which is so broad that it can refer to everything from an opinion (speak your piece) to a portion of food (a piece of pie) to a gun (a mobster’s piece) to a partner in a sexual interlude (a mobster’s piece). You begin a long meander through your mental thesaurus. “A plank used for surfing.” Better, but remember that the genus term—in the sense that you mean it—must be entered in the dictionary you’re working on. You look up “plank” to see that it generally describes a long piece of wood that is much narrower than the average surfboard. Back to the proverbial drawing board. “A panel used for surfing” is no good: “panel” in this sense usually refers to something used as a finished surface, like wood panel, or a small section of a door. “A platform used for surfing”: nope, because the pertinent meaning of “platform” here begins, “a usually raised horizontal flat surface,” which implies that there’s something underneath the platform that is raising it. But—couldn’t that be water? Perhaps, but horizontal? Are surfboards always in a horizontal orientation? What if I stick it upright into the sand—does it cease to be a platform? This line of thought smacks of overthinking; you move on. What about “a slab used for surfing”? You picture a longhair riding a tombstone.
You realize that the best genus term for “surfboard” is “board,” and you instinctively shudder with memories of Gil and “outershell.” But you aren’t using the word “surfboard” in the definition of “surfboard,” you reassure yourself; you are using the word “board.” Still, it stings a bit: it will be read as lazy, even though it’s the best genus term available. It is, after all, called a surfboard and not a surfplank or a surfslab. The only thing left is to move on to explaining how this board is different from all others.
In a standard analytical definition, the genus term is the broad category that encompasses the headword, but the differentiae are the descriptors that differentiate
each member of that category. (“Differentiae” is plural; there’s usually more than one in a definition.) “Administration,” “couch,” and “surfboard” all feature the genus term “board,” but only one of those things is used for surfing, and it’s differentiae that do that work.
Considering that a definition is made up mostly of differentiae, there’s not much in the Black Books or our training materials on the specifics of writing differentiae. There are a handful of very general suggestions and fewer hard-and-fast rules offered as guidelines, but differentiae across a dictionary can be so varied that it’s difficult to make generalizations about them. The definer is left to their own devices, primarily a knowledge of how English moves that’s based on an insatiable appetite for reading material.
The first problem you encounter is trying to decide what information to include in the differentiae: What’s vital to knowing what this word refers to, and what’s extraneous information that will at best distract and at worst confuse? In the case of a word like “surfboard,” the differentiae seem pretty clear. How is this board different from all other boards? It is, obviously, used for surfing, and this is clearly a no-brainer in terms of including this bit in your definition.
This board is also what some people would call “long” and “narrow,” though both these metrics are highly subjective (and lexicographers know that). Some people who ride what are called longboards—surfboards that are longer than the “long” you are thinking of here—consider the average surfboard stubby, and a water-skier would consider a surfboard comically wide. But you are not defining for the longboarder or the water-skier—you are defining for the mythic Everyperson, and it stands to reason that something that is as wide as an arm’s length and as tall as a person would be considered, proportionally speaking, “long” and “narrow” by Everyperson.
This board is also—and this strikes you as an important detail—buoyant, because it has to carry the weight of a person and not drag them immediately to the bottom of the sea. Shamefully, this point did not occur to you. You only noticed it because one citation in your batch is from an article called “Surfing Physics,” and it notes that surfboards are buoyant because they are less dense than the water underneath them.
You begin to scratch out a definition and place your differentiae: “a board that is long, narrow, and buoyant and which is used for surfing.” A perfectly adequate, though inelegant, definition. “Long,” “narrow,” and “buoyant” can modify “board,” and moving them before the genus gives you the tidier “a long, narrow, buoyant board used for surfing.” Congratulations: you’ve just saved twenty-five characters and one turned line. The spirit of Gove compels you to check the definition of “surf,” and you find that in our brave, new age “surf” doesn’t just refer to getting on a long, narrow, buoyant board and riding waves; it also refers to sitting on your large, spreading, squishy butt and clicking through web pages. Maybe “used for surfing” needs a little modification. “The act of surfing” could apply to either Internet or oceanic surfing, but “the sport of surfing” makes it clear you’re referring to the wave “surf” and not the web “surf.” Maybe leave “the sport of” in and let your sprachgefühl pick at it, seeing if it survives the ravages of your Germanic affliction. You are slowly getting there.
Unfortunately, you cannot slowly get there: you need to quickly get there. Lexicography in America is a ruthlessly commercial enterprise, just like the rest of publishing. Your scholarship and care mean bupkes if no one buys or uses your dictionary.
This means that the production schedule rules us with an iron fist. The Collegiate Dictionary, to take a good example, takes about two or three years to revise for a new edition. For most people, that project timeline seems ridiculously long: Three years to add some words to a dictionary? Cry me the proverbial river, lexicographers.
But a new edition doesn’t mean you’re just adding new words. The bulk of a lexicographer’s work is actually in reviewing existing entries and revising them. I didn’t draft the entry for “surfboard,” but I have reviewed it more times than I’ve actually seen a surfboard in the wild. We do the same thing in reviewing an entry as we do in writing one: we read all the citations for that word, weigh the evidence, and futz with the definition as necessary. The Collegiate Dictionary has, as of this writing, about 170,000 entries with about 230,000 definitions, give or take, to review. In reviewing, you’ll find that in the last decade or so editors have found problems with various entries and so sent pinks to those entries; you will take care of those pinks as you review and revise. Every entry, whether revised or reviewed, goes through multiple editing passes. The definer starts the job, then it’s passed to a copy editor who cleans up the definer’s work, then to a bunch of specialty editors: cross-reference editors, who make sure the definer hasn’t used any word in the entry that isn’t entered in that dictionary; etymologists, to review or write the word history; dating editors, who research and add the dates of first written use; pronunciation editors, who handle all the pronunciations in the book. Then eventually it’s back to a copy editor (usually a different one from the first round, just to be safe), who will make any additional changes to the entry that cross-reference turned up, then to the final reader, who is, as the name suggests, the last person who can make editorial changes to the entry, and then off to the proofreader (who ends up, again, being a different editor from the definer and the two previous copy editors). After the proofreaders are done slogging through two thousand pages of four-point type, the production editors send it off to the printer or the data preparation folks, and then we get another set of dictionary pages (called page proofs) to proofread.
This process happens continuously as we work through a dictionary, so a definer may be working on batches in C, cross-reference might be in W, etymology in T, dating and pronunciation in the second half of S, copy editors in P (first pass) and Q and R (second pass), while the final reader is closing out batches in N and O, proofreaders are working on M, and production has given the second set of page proofs to another set of proofreaders for the letter L. We all stagger our way through the alphabet until the last batch, which is inevitably somewhere near G, is closed. By the time a word is put in print either on the page or online, it’s generally been seen by a minimum of ten editors. Now consider that when it came to writing the Collegiate Dictionary, Eleventh Edition, we had a staff of about twenty editors working on it: twenty editors to review about 220,000 existing definitions, write about 10,000 new definitions, and make over 100,000 editorial changes (typos, new dates, revisions) for the new edition. Now remember that the 110,000-odd changes made were each reviewed about a dozen times and by a minimum of ten editors. The time given to us to complete the revision of the Tenth Edition into the Eleventh Edition so production could begin on the new book? Eighteen months.
You can see, then, why these definition goofs happen. Is it any wonder that when faced with that pressure, one editor working his way through the Third and toward an early grave defined “fishstick” as “a stick of fish”? You can almost see the oily sheen of desperation on the page: there, this batch is done, I’m done, get me out of here.
A definition like the one at “fishstick” seems incongruous when compared with other definitions in the Third. Unabridged dictionaries don’t just include more entries but longer definitions with more complex differentiae than are found in abridged dictionaries. An unabridged dictionary gives a lexicographer the space to stretch out a little bit, because users expect to see more of everything in an unabridged dictionary. But this leads to problems: When space isn’t an issue, how do you know when to stop?
It was evidently very easy to lose sight of the forest for all that delicious xylem and phloem. This is one (now infamous) definition of “hotel” found in Webster’s Third:
a building of many rooms chiefly for overnight accommodation of transients and several floors served by elevators, usually with a large open street-level lobby containing easy chairs, with a variet
y of compartments for eating, drinking, dancing, exhibitions, and group meetings (as of salesmen or convention attendants), with shops having both inside and street-side entrances and offering for sale items (as clothes, gifts, candy, theater tickets, travel tickets) of particular interest to a traveler, or providing personal services (as hairdressing, shoe shining), and with telephone booths, writing tables and washrooms freely available
Dancing compartments! Travel tickets! Candy! Where are these hotels today?
Lexicographers tend to fall into one of two categories when it comes to writing definitions: lumpers and splitters. Lumpers are definers who tend to write broad definitions that can cover several more minor variations on that meaning; splitters are people who tend to write discrete definitions for each of those minor variations. This seems to be a natural inclination: lumpers have a very hard time teasing out the micro-meanings covered by their broad definitions, and splitters have a very hard time collapsing their incredibly concise definitions into one. Emily and I are splitters; Neil is a lumper; Steve can do both, though the more he defines, the more he tends toward lumping. That definition of “fishstick” was written by a lumper; that definition of “hotel” was definitely written by a splitter.